Surgery for temporal lobe epilepsy in children: relevance of presurgical evaluation and analysis of outcome

OBJECT: The authors' goal in this paper was to retrospectively evaluate the relevance of the presurgical workup and the postoperative outcome in children (< 15 years) who undergo surgery for temporal lobe epilepsy (TLE). METHODS: The authors performed a retrospective analysis of 68 patients (43 boys and 25 girls) who underwent resection for TLE between 2001 and 2010 at a single center and had a minimum postoperative follow-up of 12 months. Presurgical investigations included full clinical evaluation, interictal electroencephalography (EEG), and MRI in all cases; cognitive evaluation in patients older than 5 years; scalp video-EEG in 46 patients; and invasive EEG in 3 patients. Clinical evaluation included a careful assessment of ictal semiology (based on anamnestic reports or video-EEG review), with particular attention to early signs and/or symptoms suggestive of temporal lobe origin of the seizure. Microsurgical resections were performed within the anatomical limits of the temporal lobe, and surgical specimens were processed for histological examination. Postoperative assessment of seizure outcome (Engel classification system) and cognitive performance was conducted at regular intervals. The effect on postoperative seizure outcome (good = Engel Class I; poor = Engel Classes II-IV) of several presurgical and surgical variables was investigated by bivariate statistical analysis. RESULTS: All patients had at least 1 early sign or symptom suggesting a temporal lobe origin of their seizures. Lateralized interictal or ictal EEG abnormalities were seen in all patients, and they were localized to the temporal lobe in 45 patients. In all cases MRI demonstrated a structural abnormality. Surgery consisted of a tailored anterior temporal lobectomy in 64 patients and a neocortical lesionectomy in 4 patients. Postoperatively, 58 patients (85%) were in Engel Class I. Variables significantly associated with a poor outcome were preoperative sensory motor deficit (p = 0.019), mental retardation (p = 0.003), MRI abnormalities extending outside the temporal lobe (p = 0.0018), history of generalized seizures (p = 0.01) or status epilepticus (p = 0.008), unremarkable histology (p = 0.001), seizures immediately postoperatively (p = 0.00001), and ipsilateral epileptiform activity on postoperative EEG (p = 0.005). At postoperative neuropsychological assessment, the percentage of patients with a pathological score at the final visit invariably decreased compared with that at the preoperative evaluation in all considered cognitive domains. CONCLUSIONS: Among the study population, a surgical selection based on a noninvasive evaluation was possible in most patients. The invaluable information resulting from the rigorous noninvasive electroclinical and neuroimaging evaluation can lead to excellent surgical results without the use of invasive, time-consuming, and expensive diagnostic tools. The potential reduction of invasiveness-related risks, complexity, and costs of presurgical investigations should hopefully allow for an increase in the number of children with TLE who will receive surgery, particularly in centers with limited technological resources.